JP2008219575A - Solid-state imaging device inspection method, solid-state imaging device inspection apparatus, and solid-state imaging device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a solid-state imaging device inspection method which appropriately evaluates the quality of a solid-state imaging device by inspecting it with sufficient accuracy based on characteristics of peculiar pixels. <P>SOLUTION: The solid-state imaging device is determined whether it is acceptable as a product or not through first to third inspection processes. In the first inspection process, light is blocked from reaching the solid-state imaging device to acquire image data. In the second inspection process, brightness data of the isolated peculiar pixels is removed from the image data and close-packed peculiar pixels consisting of a set of several peculiar pixels is detected. In the third inspection process, the solid-state imaging device is inspected per predetermined range set in advance, and the number of peculiar pixels within each range is counted, and at the same time, the sum of brightness data of the individual peculiar pixels within each range is calculated as a defect level. The quality of the solid-state imaging device to be inspected is determined based on the magnitude of the defect level. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a method and apparatus for inspecting a solid-state image sensor, and more particularly to a method and apparatus for inspecting a pixel defect of a solid-state image sensor.

  Imaging devices that use a solid-state imaging device to acquire subject images as digital images, such as digital cameras, video cameras, mobile phones with cameras, and telescopes, are widely used. In recent years, with the widespread use of these imaging devices, solid-state imaging devices are required to have high sensitivity, high pixels, and low cost.

  As solid-state imaging devices, CCDs, CMOSs, and the like are known, and all are manufactured by finely processing a semiconductor such as Si. Although a manufacturing method of such a solid-state image sensor has been established at present, as the sensitivity of the solid-state image sensor increases and the number of pixels increases, the defects of the pixels of the solid-state image sensor become relatively obvious and imaged. This is an obstacle to improving the image quality and reducing the price of the solid-state imaging device.

  For example, it is known that sensitivity unevenness caused by substrate unevenness that occurs during semiconductor processing and dark current caused by thermally excited charges appear in the image as fixed pattern noise. In addition, it is known that crystal defects of a semiconductor serving as a base material appear in an image as so-called white scratches. Further, since this white scratch is caused by a semiconductor crystal defect, it is known that the white scratch occurs suddenly due to the influence of cosmic rays or the like even after the solid-state imaging device is manufactured, and increases with time.

  In practice, it is extremely difficult to completely remove such defects of the solid-state imaging device that cause image quality degradation. Therefore, in order to improve the yield and provide a solid-state imaging device at a low cost, a pixel that outputs unique data (hereinafter referred to as a unique pixel) is detected, and depending on the number of unique pixels or the number of unique pixels. Thus, the quality of the solid-state image sensor is determined. Therefore, a solid-state imaging device including a specific pixel and a set of specific pixels to some extent is generally distributed.

Also, with respect to white scratches that increase with time, a method of easily detecting white scratches has been proposed in order to correct this by image processing (see, for example, Patent Document 1).
JP 2002-314055 A

  As described above, the quality of the solid-state imaging device is determined according to the number of singular pixels or the number of sets of singular pixels. However, it is difficult to say that the inspection of the solid-state imaging device performed based on the number of singular pixels or the number of singular pixel sets is sufficiently accurate.

  That is, even if the number of singular pixels or the number of singular pixels is less than or equal to a predetermined number and is sufficiently small, for example, if these are singular pixels that always output extremely large data, In the captured image, it appears as a conspicuous high-brightness point, and the deterioration of the image quality is serious.

  In addition, for example, if the singular pixel can be corrected based on the data of the surrounding pixels, or if it is a singular pixel that outputs data that is not much different from normal data, the image quality degradation is slight, and general use The target digital camera or the like may be sufficiently usable.

  In this way, the solid-state imaging device inspection method that detects the number of singular pixels or a set of singular pixels and uses them alone to determine the quality of the solid-state imaging device is sufficient for the characteristics of the singular pixels and the set of singular pixels. There is a problem that it is not reflected and the inspection is not sufficiently accurate.

  In addition, such an inaccurate inspection may cause a defective solid-state image sensor to be a non-defective product. Therefore, the reliability of the defective solid-state image sensor may be impaired by the distribution of such a defective solid-state image sensor. Further, the inspection with insufficient accuracy may determine that a solid-state image sensor that should be a good product is defective, and as a result, the yield is deteriorated and the price of the solid-state image sensor is hindered.

  The present invention has been made in view of the above-described problems. An inspection method and an inspection apparatus for appropriately detecting and evaluating a specific pixel or a set of specific pixels and inspecting a defect of a solid-state imaging device with sufficient accuracy are provided. The purpose is to provide.

  According to the inspection method of the present invention, a first inspection step of acquiring image data in a state where a solid-state image sensor to be inspected is shielded, and data output from each pixel of the solid-state image sensor is compared with a threshold based on the image data. And a second inspection step for detecting a singular pixel that outputs data having a value larger than the threshold value, and data that is inspected in units of a predetermined range of the image data and is output by the singular pixel within the predetermined range. And a third inspection step for determining whether the solid-state imaging device is good or bad by calculating a defect level that is a total value of the above and comparing the defect level with a predetermined upper limit value.

  In the third inspection step, the defect level is calculated, the number of the specific pixels within the range to be inspected is counted, and the pass / fail determination performed by comparing the defect level with a predetermined upper limit value; The quality determination of the solid-state imaging device is performed in combination with the quality determination performed according to the number of unique pixels.

  In addition, according to the result of the quality determination of the solid-state image sensor performed according to the number of the specific pixels, the quality determination of the solid-state image sensor performed by comparing the defect level with a predetermined upper limit value is performed. It is characterized by that.

  In addition, when the solid-state imaging device is determined to be defective in the quality determination of the solid-state imaging device performed according to the number of the singular pixels, the solid state is performed by comparing the defect level with a predetermined upper limit value. It is characterized in that the quality of the image sensor is determined.

  In addition, when the solid-state image sensor is determined to be good in the quality determination of the solid-state image sensor performed according to the number of singular pixels, the solid state is performed by comparing the defect level with a predetermined upper limit value. It is characterized in that the quality of the image sensor is determined.

  In addition, according to the result of the quality determination of the solid-state image sensor performed by comparing the defect level with a predetermined upper limit value, the quality determination of the solid-state image sensor performed according to the number of the specific pixels is performed. It is characterized by that.

  In the second inspection step, the singular pixel is detected, and the data of the isolated singular pixel, which is a pixel in which all adjacent pixels output data having a value equal to or less than the threshold, is removed from the image data. Inspection image data in which only the dense pixel data is left is created, and the third inspection step is performed based on the inspection image data.

  The number of pixels included in the predetermined range is 0.001% to 0.01% of all pixels of the solid-state imaging device.

  The inspection apparatus of the present invention compares data output from each pixel of the solid-state image sensor with a threshold value based on image data obtained by imaging the solid-state image sensor to be inspected in a light-shielded state, and is larger than the threshold value. Singular pixel detection means for detecting singular pixels that output value data, and a defect level that is a total value of data output by the singular pixels within the predetermined range in units of the predetermined range of the image data. And a defect level evaluation unit that performs pass / fail determination by comparing the defect level with a predetermined upper limit value.

  In addition, a singular pixel counting unit that counts the number of the singular pixels within the predetermined range and a singular pixel number evaluation unit that performs a pass / fail determination according to the number of the singular pixels are provided, and the pass / fail by the defect level evaluation unit is provided. The determination of the quality of the solid-state imaging device is performed by combining the determination and the quality determination by the unique pixel number evaluation unit.

  In addition, the pass / fail determination by the defect level evaluation unit is performed according to a result of the pass / fail determination by the singular pixel number evaluation unit.

  In addition, the quality determination by the defect level evaluation unit is performed when the solid-state imaging device is determined to be defective as a result of the quality determination by the specific pixel number evaluation unit.

  In addition, the pass / fail determination by the defect level evaluation unit is performed when the solid-state imaging device is determined to be good as a pass / fail determination defect by the singular pixel number evaluation unit.

  Further, the pass / fail determination by the singular pixel number evaluating means is performed according to the result of the pass / fail determination by the defect level evaluating means.

  The solid-state image sensor of the present invention is characterized by being inspected by the above-described solid-state image sensor inspection method.

  The solid-state image sensor of the present invention is characterized by being inspected by the above-described solid-state image sensor inspection apparatus.

  According to the solid-state image sensor inspection method and the solid-state image sensor inspection apparatus of the present invention, the solid-state image sensor can be inspected with sufficient accuracy and appropriately evaluated based on the characteristics of defective pixels of the solid-state image sensor.

  As shown in FIG. 1, an inspection apparatus (solid-state image sensor inspection apparatus) 10 that inspects a solid-state image sensor includes a light shielding mechanism 21 that arranges a CCD (solid-state image sensor) 15 to be inspected. The light shielding mechanism 21 shields the CCD 15 connected to the inspection apparatus 10 from light in a wavelength region having sensitivity of the CCD 15 such as visible light and infrared light in order to perform defect inspection.

  The CCD 15 is a so-called solid-state imaging device that photoelectrically converts light received on the light receiving surface and outputs an analog imaging signal. The light receiving surface of the CCD 15 is provided with, for example, rectangular pixels arranged on one surface, and photoelectric conversion is performed for each pixel. That is, each pixel accumulates a charge proportional to the amount of received light and outputs this as an imaging signal. The operation of the CCD 15 to be inspected is controlled by the CPU 23 via the CCD driver 22.

  The imaging signal output from the CCD 15 is input to a correlated double sampling circuit (CDS) 24, noise generated by reading the imaging signal is removed, and the signal amplification circuit (AMP) 26 amplifies the imaging signal. The amplified imaging signal is converted into digital image data by an A / D converter (A / D) 27.

  This image data is image data that is exactly proportional to the amount of charge stored in each pixel of the CCD 15 and is input to a DSP (Digital Signal Processor) 28. The DSP 28 includes an image input controller 31, a YC conversion processing circuit 32, a compression / decompression processing circuit 33, and the like. The DSP 28 temporarily stores image data in the RAM 34 and performs various image processing on the image data.

  The image input controller 31 buffers the image data input from the A / D 27 and stores it in the RAM 34 connected via the data bus 36. The YC conversion processing circuit 32 reads image data from the RAM 34 and converts it into a luminance signal Y and color difference signals Cr and Cb. The compression / decompression processing circuit 33 compresses the YC converted image data by a method such as TIFF or JPEG, and outputs the compressed data in a predetermined file format.

  The RAM 34 is a working memory, temporarily stores image data, and is loaded with a control program for the inspection apparatus 10 executed by the CPU 23. The ROM 37 stores a control program for the inspection apparatus 10 and various setting information. The various setting information stored in the ROM 37 includes, for example, an exposure time when imaging using the CCD 15 to be inspected, a threshold value used when detecting a specific pixel, a predetermined range for evaluating the characteristic of the specific pixel of the CCD 15, There are an upper limit number of singular pixels allowed in the range, a target number of singular pixels within a predetermined range, which is a guide for a good CCD 15, and a value to be compared with the total value of luminance data of the singular pixels within the predetermined range. Note that the inspection apparatus 10 inspects the CCD 15 by changing the inspection conditions according to the type of the CCD 15 to be inspected and the like, not limited to the values of various settings stored in the ROM 37.

  Substantial inspection of the CCD 15 by the inspection apparatus 10 is performed by the inspection unit 41. The inspection unit 41 (quality determination means) detects the specific pixels of the CCD 15 based on the image data captured by the CCD 15 to be inspected, and determines the quality of the CCD 15 as a product. The inspection unit 41 includes a singular pixel detection unit 42, an isolated singular pixel removal unit 43, a singular pixel characteristic calculation unit 44, a singular pixel number evaluation unit 45, a defect level evaluation unit 46, and the like.

  The singular pixel detection unit 42 (single pixel detection means) reads out the image data stored in the RAM 34 and compares the luminance data of each pixel of the image data with a predetermined threshold value of the luminance data. Then, a pixel having luminance data exceeding the threshold is detected as a specific pixel, and the number of the specific pixels is counted. Further, the singular pixel detection unit 42 is configured so that, for example, about 20,000 pixels of luminance data out of image data of 3.6 million pixels is larger than the threshold value so that a predetermined number of pixels are detected as singular pixels. Thus, the threshold value is adjusted, and the adjusted threshold value is stored in the RAM 34 as an effective threshold value.

  The isolated singular pixel removing unit 43 investigates whether or not the pixels adjacent to each of about 20,000 singular pixels detected by the singular pixel detection unit 42 are singular pixels based on the image data and the effective threshold. If all the pixels adjacent to the singular pixel are not singular pixels, the singular pixel is determined to be an isolated singular pixel, and the luminance data of the isolated singular pixel is removed. The isolated unique pixel removing unit 43 stores the image data from which the brightness data of the isolated unique pixels is removed in the RAM 34 as inspection image data. One or more other singular pixels always exist around the singular pixels remaining in the inspection image data.

  The isolated singular pixel removing unit 43 stores the coordinates of the singular pixel from which the luminance data is removed in the RAM 34 as isolated singular pixel coordinate data. This isolated pixel coordinate data is used when interpolating and correcting the brightness data of the isolated singular pixel based on the brightness data of the normal pixel adjacent to each isolated singular pixel in an individual imaging device using the inspected CCD 15. .

  The singular pixel characteristic calculation unit 44 (defect level calculation unit, singular pixel counting unit) reads the inspection image data and inspects the characteristic of the singular pixel of the CCD 15. Specifically, the singular pixel characteristic calculation unit 44 counts the number of singular pixels within a predetermined range as the characteristic of the singular pixel of the CCD 15. Further, the singular pixel characteristic calculation unit 44 calculates the sum of the luminance data of the singular pixels within a predetermined range as the characteristic of the singular pixel and sets this as the defect level.

  The singular pixel number evaluation unit 45 (single pixel number determination means) compares the number of singular pixels within the predetermined range counted by the singular pixel characteristic calculation unit 44 with the predetermined number to evaluate the quality of the CCD 15 as a product. To do. As the predetermined number used for comparison with the number of singular pixels, for example, a target number that is a target of a non-defective product, and an allowable upper limit number that is an upper limit number of singular pixels allowed to be included in a predetermined range, Is predetermined.

  The defect level evaluation unit 46 (defect level evaluation means) compares the defect level within the predetermined range calculated by the singular pixel characteristic calculation unit 44 with the predetermined level, and evaluates the quality of the CCD 15 as a product. Note that, as the predetermined level used for comparison with the defect level, a specified level stored in the ROM 37 is used, or an effective level determined according to an effective threshold stored when detecting a singular pixel is used. When the effective level is used as the predetermined level, the defect level evaluation unit 46 calculates the effective level based on the specified level and the effective threshold value.

  The operation of the inspection apparatus 10 configured as described above will be described. As shown in FIG. 2, when the CCD 15 to be inspected is arranged in the light shielding mechanism 21 and the inspection of the CCD 15 is executed, first, a first inspection step is performed in which the CCD 15 takes an image and obtains image data. Specifically, the inspection apparatus 10 transfers the charges already accumulated in the pixels of the CCD 15 and discards this imaging signal. Immediately after discarding the imaging signal in this way, exposure is performed in the light shielding mechanism 21 for a predetermined time, and image data is obtained from the imaging signal. Since this image data is image data exposed in the light shielding mechanism 21, ideally the luminance data of all the pixels is zero. However, the image data obtained in the first inspection process actually includes pixels whose luminance data is not 0 due to fixed pattern noise or white scratches.

  Then, as described above, the second inspection process for detecting the specific pixels of the CCD 15 based on the image data obtained in the first inspection process is performed. In the second inspection step, the luminance data of each pixel is compared with a predetermined threshold value, and a specific pixel having luminance data exceeding the threshold value is detected. At this time, when the number of singular pixels is less than a predetermined number, for example, 20,000 out of 6.3 million pixels, detection of singular pixels is repeatedly performed using an effective threshold value smaller than the predetermined threshold value. .

  Thus, when a predetermined number or more of unique pixels are detected, isolated singular pixels whose adjacent pixels are not singular pixels are detected. Since the luminance data of the isolated singular pixel is interpolated and correctly corrected based on the luminance data of the adjacent pixel, it is not a serious defect. Therefore, the luminance data of the isolated singular pixels is removed from the image data, and inspection image data is created in which only the luminance data of a portion where a plurality of singular pixels are dense is left.

  As shown in FIG. 3, in the third inspection step, the quality of the CCD 15 as a product is determined based on the inspection image data. In the third inspection step, inspection is performed for each predetermined range of inspection image data, for example, a region of 16 × 16 pixels. That is, the number of singular pixels within a predetermined range for executing the inspection is counted. At the same time, the sum of the luminance data of the singular pixels within the predetermined range to be inspected is calculated, and a defect level serving as a guide indicating the degree of the defect within the predetermined range is obtained.

  This defect level is compared with a predetermined specified level (upper limit value) to determine whether the CCD 15 is good or bad. The specified level used here is a value that is accepted as a total value of the luminance data of the peculiar pixels included in the non-defective CCD 15 within a predetermined range, and is a value set in advance before the start of the inspection.

  Accordingly, when the defect level in the predetermined range is a value equal to or higher than the specified level, the CCD 15 is determined to be defective as a product, and the inspection of the CCD 15 ends.

  On the other hand, when the defect level in the predetermined range is smaller than the specified level, the peculiar pixels in the predetermined range inspected are allowed. That is, the CCD 15 is determined to be a non-defective product within the predetermined range inspected. The region to be inspected is moved by one pixel in the vertical (or horizontal) direction of the inspection image (see the arrow in FIG. 4B), and the inspection is repeated in the above-described procedure. In addition, when it is determined that the entire area of the inspection image data is a non-defective product in the above-described inspection, the inspected CCD 15 is finally determined to be a non-defective product, and the inspection is completed.

  Specifically, for example, as shown in FIG. 4A, when imaging is performed while the CCD 15 to be inspected is shielded from light, a large number of point images are captured in the obtained image data 51. These point images are all due to pixel defects such as white scratches and fixed pattern noise. Among these point images, the brightness data of isolated singular pixels can be easily corrected.

  Therefore, as shown in FIG. 4B, inspection image data 52 is created by removing the luminance data of isolated singular pixels from the image data 51, and this inspection image data 52 is used to determine whether the CCD 15 is good or bad. The point image remaining in the inspection image data 52 is due to the presence of two or more specific pixels densely. Since the luminance data of such dense singular pixels includes singular pixels in adjacent pixels, the data cannot be appropriately interpolated and corrected based on the luminance data of surrounding pixels. The degree of influence of such a set of singular pixels on an image is estimated for each predetermined range 53. That is, the number of unique pixels in the predetermined range 53 is counted, and the total value of the luminance data of these unique pixels is calculated as the defect level.

  For example, as shown in FIG. 4C, it is assumed that the dense singular pixel 54 is one set of four singular pixels, and only the defect 54 is included in the predetermined range 53. Of the four singular pixels constituting the dense singular pixel 54, two singular pixels output extremely large luminance data as compared with the threshold value, and the other two singular pixels are larger than the threshold value but have a threshold value. It is assumed that the pixel is a singular pixel that outputs luminance data of approximately the same level as. Such a dense singular pixel 54 is actually a very conspicuous defect because there are a small number of singular pixels, but singular pixels that output large luminance data are densely present.

  When inspecting the predetermined range 53 including such dense singular pixels 54, if the quality of the CCD 15 is determined simply based only on the number of singular pixels, the predetermined range of the CCD 15 to be inspected is determined from the small number of singular pixels. 53 is judged to be a non-defective product. However, the inspection apparatus 10 compares the defect level of the predetermined range 53 including the dense singular pixels 54 with the specified level, and if the defect level is larger than the specified level, regardless of the number of singular pixels, It is determined that the CCD 15 to be inspected contains a serious defect, and the CCD 15 is correctly determined as defective.

  Further, for example, as shown in FIG. 4D, it is assumed that the predetermined range 53 to be inspected includes dense singular pixels 56a, 56b, 56c, and 56d composed of a set of several singular pixels. Further, the singular pixels constituting the dense singular pixels 56a, 56b, 56c, and 56d are all singular pixels that are larger than the threshold value but output luminance data substantially equal to the threshold value. Such dense singular pixels 56a, 56b, 56c, and 56d are all included in the predetermined range 53. Therefore, although the number of singular pixels is large as a whole, the luminance data output from them is the luminance data of normal pixels. There is no big difference, and depending on the type of the image pickup apparatus using the CCD 15 or the like, it may not be a big problem.

  When the predetermined range 53 including such dense singular pixels 56a, 56b, 56c, and 56d is inspected, if the quality of the CCD 15 is determined simply based on the number of singular pixels, the inspection is performed based on the number of singular pixels. The predetermined range 53 of the CCD 15 to be determined is determined to be defective, and as a result, the CCD 15 is determined to be defective. However, the inspection apparatus 10 compares the defect level in the predetermined range 53 including the dense singular pixels 56a, 56b, 56c, and 56d with the specified level, and if the defect level is smaller than the specified level, the number of defective pixels is determined. In spite of the large number, it is determined that the CCD 15 to be inspected does not contain a serious defect, and the predetermined range 53 of the CCD 15 is correctly determined as a non-defective product.

  As described above, when the defect of the pixel is inspected, if the quality of the CCD 15 is determined based on the defect level, it is possible to accurately determine the quality of the CCD 15 with sufficient consideration of the characteristics of the specific pixels. Further, as a result of accurately determining whether the CCD 15 is good or bad, the reliability of the product is increased. Furthermore, since the yield is improved, the price of the CCD 15 can be reduced.

  As described above, when a pixel defect is inspected, the quality of the CCD 15 may be determined based on the defect level, and the quality of the CCD 15 may be determined based on the number of unique pixels.

  For example, as shown in FIG. 5, image data obtained by shielding light using a CCD 15 inspected in the first inspection step is obtained. In the second inspection step, the brightness data of isolated singular pixels is removed from the image data, and inspection image data is created. In the third inspection step in which the quality of the CCD 15 is actually determined based on the inspection image data, the quality of the CCD 15 is determined from the number of specific pixels, and the quality of the CCD 15 is determined from the defect level. Pass / fail is determined.

  Specifically, first, the number of unique pixels in the predetermined range 53 to be inspected is counted, and at the same time, the defect level in the predetermined range 53 is calculated. Then, the counted number of unique pixels in the predetermined range 53 is compared with the target number. As described above, the predetermined target number is the number of target specific pixels included in the predetermined range 53 by the high-quality CCD 15. Therefore, if the number of singular pixels in the predetermined range 53 to be inspected is less than the target number, the predetermined range 53 of the CCD 15 is determined to be good, and the region to be inspected is moved.

  On the other hand, when the number of singular pixels in the predetermined range 53 is larger than the target number, the number of singular pixels in the predetermined range 53 is compared with the allowable upper limit number. As described above, the predetermined allowable upper limit number is the upper limit number allowed to be included in the predetermined range 53. Therefore, when the number of specific pixels in the predetermined range 53 to be inspected is larger than the allowable upper limit number, the predetermined range 53 of the CCD 15 is determined to be defective, and as a result, the CCD 15 is determined to be defective as a product.

  That is, even if the degree of degradation of the image quality of individual singular pixels is not so great, if many singular pixels are included, the overall image becomes rough. If a singular pixel is included, the CCD 15 is determined to be defective. When the number of unique pixels in the predetermined range 53 is larger than the target number and smaller than the allowable upper limit number, the quality determination based on the defect level is performed as described above.

  As described above, when the number of singular pixels in the predetermined range 53 and the defect level in the predetermined range 53 are combined to determine whether the CCD 15 is good or bad, the characteristics of the singular pixels are sufficiently taken into account to perform an accurate inspection. it can.

  For example, when the number of singular pixels larger than the target number is included in the predetermined range 53, if the quality is simply determined based on the number of singular pixels, all of the CCDs 15 should be non-defective products without serious defects. It is determined to be defective, and the yield is very poor.

  However, as described above, not only the pass / fail determination based on the target number, but also the pass / fail determination based on the allowable upper limit count and the pass / fail determination based on the defect level are combined to perform the characteristics of the specific pixels in the predetermined range 53. A quality determination with sufficient consideration is made. That is, among the CCDs 15 that are uniformly judged as defective in the quality determination based on the target number of specific pixels, the CCD 15 that does not include a significant characteristic and can withstand its original use is correctly determined as a good product.

  For example, if only the quality determination based only on the defect level is performed, the CCD 15 that generates an image that is rough over a wide range is determined as a non-defective product when the degree of deterioration of the image quality of each individual pixel is not so large. Sometimes. However, as described above, in addition to the pass / fail determination based on the defect level, the pass / fail determination based on the number of defective pixels is performed, so that the pass / fail determination with sufficient consideration of the characteristics of the defect within the predetermined range 53 is performed. That is, although the defect level is not so high, the CCD 15 that generates a rough image as a whole is correctly determined to be defective.

  As described above, the quality determination based on the number of singular pixels and the quality determination based on the defect level are performed in combination, but the number of singular pixels in the predetermined range 53 to be inspected is compared with the target number, and the CCD 15 There is no need to make a pass / fail judgment.

  For example, as shown in FIG. 6, first, image data obtained by shading using the CCD 15 to be inspected in the first inspection step as described above is obtained. In the second inspection step, the brightness data of isolated singular pixels is removed from the image data, and inspection image data is created. The quality of the CCD 15 is actually determined based on the inspection image data. In the third inspection step, the quality of the CCD 15 is determined from the number of unique pixels, the quality of the CCD 15 is determined from the defect level, and the quality of the CCD 15 is determined.

  In the third inspection step, the number of singular pixels in the predetermined range 53 is compared with the allowable upper limit number. If the number of unique pixels in the predetermined range 53 is larger than the allowable upper limit number, the CCD 15 is determined to be defective. On the other hand, when the number of singular pixels within the predetermined range 53 is smaller than the allowable upper limit number, the quality of the CCD 15 is judged based on the defect level.

  As described above, the pass / fail judgment based on the size comparison between the number of singular pixels and the allowable upper limit number and the pass / fail judgment based on the defect level are based on the fact that the number of singular pixels included in the predetermined range 53 is larger than the target number. Even if the number is small, it is substantially the same as performing the pass / fail determination based on the defect level.

  Therefore, according to the pass / fail determination based only on the number of singular pixels, it is determined as defective due to the large number of singular pixels even though it is a CCD 15 that does not include a serious defect and can withstand practical use. The resulting CCD 15 is correctly determined as a non-defective product.

  Furthermore, even if the CCD 15 includes a serious defect and cannot withstand practical use, if only the pass / fail determination based on the number of unique pixels is performed, the product is determined to be a non-defective product due to the small number of unique pixels. The CCD 15 is correctly determined to be defective.

  As described above, when the quality determination of the CCD 15 and the quality determination based on the defect level in the predetermined range 53 are performed in combination based on the size comparison between the number of specific pixels in the predetermined range 53 and the allowable upper limit number, These characteristics are fully taken into account, and an accurate inspection can be performed.

  In the above example, when the determination of pass / fail of the CCD 15 based on the number of singular pixels and the pass / fail determination of the CCD 15 based on the defect level are performed in combination with the above example, the pass / fail determination based on the number of singular pixels is performed first. Accordingly, the quality determination of the CCD 15 based on the defect level is performed, but not limited to this, the quality determination based on the defect level is performed first, and the quality determination of the CCD 15 is performed based on the number of defective pixels according to the result. good.

  For example, as shown in FIG. 7, when the defect level within the predetermined range 53 is larger than the specified level, the CCD 15 is determined to be defective without waiting for a pass / fail determination based on the number of unique pixels. On the other hand, when the defect level in the predetermined range 53 is smaller than the specified level, the number of unique pixels in the predetermined range 53 is compared with the allowable upper limit number. Even if the defect level in the predetermined range 53 is smaller than the specified level, if the number of unique pixels in the predetermined range 53 is larger than the allowable upper limit number, the CCD 15 is determined to be defective. Further, when the defect level in the predetermined range 53 is smaller than the specified level and the number of unique pixels in the predetermined range 53 is smaller than the allowable upper limit number, the predetermined range 53 of the CCD 15 is determined as a non-defective product.

  In this way, even when the pass / fail judgment of the CCD 15 based on the defect level is performed first, and the pass / fail judgment of the CCD 15 based on the number of singular pixels is made according to the result, the singular pixel of the CCD 15 is the same as described above. It is possible to perform a highly accurate inspection with sufficient consideration of the above characteristics.

  In the above-described example, when the quality of the CCD 15 is determined based on the defect level, the defect level in the predetermined range 53 is compared with a predetermined specified level. The pass / fail judgment of the CCD 15 may be performed by comparing the effective level (upper limit value) determined according to the threshold and the defect level within the predetermined range 53.

  As described above, the prescribed level as a rough standard for the defect level is determined in advance according to the product specifications and the like. However, for singular pixels that generate fixed pattern noise, white scratches, etc., the brightness data value that is output changes according to the temperature at the time of inspection, and even the same singular pixel is output depending on various conditions such as the environment at the time of inspection. The value of the brightness data is different. Therefore, the inspection apparatus 10 adjusts a threshold value used when detecting a singular pixel so that a predetermined number of pixels are detected as singular pixels, and whether or not the detected singular pixel is a serious defect. Is determined based on the defect level and the number of unique pixels.

  As described above, the threshold value serving as a reference for determining whether or not each pixel is a specific pixel is adjusted so that the accuracy of the inspection of the CCD 15 does not deteriorate due to factors such as the environment at the time of inspection. The reference value to be compared with the defect level when determining whether or not the CCD 15 is good is not necessarily a predetermined specified level, and may be an effective level determined according to an effective threshold.

  For example, when determining whether the CCD 15 is good or bad based on the defect level, comparing the predetermined specified level with the defect level, the CCD 15 is likely to be determined to be defective depending on the environment at the time of inspection, or vice versa. In some cases, the CCD 15 is likely to be judged as good. This is due to determining whether each pixel is a singular pixel using an execution threshold value that is adjusted so that a predetermined number of pixels are detected as singular pixels. Accordingly, in order to make the inspection accuracy constant regardless of the environment at the time of inspection, when performing the pass / fail determination of the CCD 15 based on the defect level, an effective level determined according to the effective threshold based on the specified level, It is preferable to compare the defect level.

  As described above, when the quality of the CCD 15 is determined based on the defect level, if the quality of the CCD 15 is determined by comparing the effective level with the defect level, the accuracy of the CCD 15 can be determined with high accuracy regardless of the environment during the inspection. Inspection can be performed.

  In the above-described embodiment, as an example of the predetermined range which is a unit for performing the inspection, in the case of the CCD 15 having 63,000,000 pixels, the inspection is performed in units of 16 × 16 pixels, but is not limited thereto. For example, if the predetermined range for performing the inspection is too small, the inspection accuracy and the inspection speed are deteriorated, and if the predetermined range for performing the inspection is too large, the inspection speed may be remarkably deteriorated. Therefore, it is preferable that the inspection is performed by setting a range including 0.001% or more and 0.01% or less of the total number of pixels of the solid-state imaging device to be inspected as a predetermined range. More preferably, it is preferable to inspect a range including pixels of 0.002% or more and 0.008% or less as a predetermined range, and particularly preferably, a range including pixels of 0.003% or more and 0.006% or less is predetermined. It is preferable to inspect as a range. By inspecting such an appropriate range as a predetermined range, it is possible to inspect the solid-state imaging device quickly and with high accuracy. In addition, a value that is a criterion for pass / fail judgment such as a specified level must be changed appropriately according to the size of the predetermined range.

  In the above-described embodiment, the imaging signal from the CCD 15 to be inspected is converted into digital image data, and the CCD 15 is inspected by image processing. However, the present invention is not limited to this, and the imaging signal is not converted into digital image data. The CCD 15 may be inspected in the same manner as in the above-described embodiment by using an analog imaging signal and electrically processing this.

  In the above-described embodiment, a CCD is described as an example of a solid-state imaging device to be inspected. However, the present invention is not limited to this, and other solid-state imaging devices such as a CMOS can also be inspected similarly to the above-described embodiment.

  In the above-described embodiment, since the correction is easily performed, the CCD 15 is inspected using the inspection image data obtained by removing the luminance data of the isolated singular pixels from the image data. The CCD 15 may be inspected using image data from which luminance data has not been removed.

  In the above-described embodiment, the inspection of the predetermined range of the inspection image data is performed to determine whether the CCD 15 is good or bad. However, the present invention is not limited to this. Alternatively, the defect level for each predetermined range may be stored in a table format or the like, and the quality of the CCD 15 may be determined based on this.

  Furthermore, in the above-described embodiment, whether the CCD 15 is good or bad is determined according to the characteristics of the singular pixels in the predetermined region. However, the present invention is not limited to this. And the defect level for each dense specific pixel may be calculated, and the quality of the CCD 15 may be determined based on the characteristics of each dense specific pixel.

  In the above-described embodiment, fixed pattern noise and white scratches are described as an example of defects of the CCD 15. However, the present invention is not limited to this, and defects that affect the imaging signal obtained from the CCD 15 or image data based thereon. If it is present, it can be easily detected by the method and apparatus of the present invention, and the CCD 15 can be accurately inspected.

It is a block diagram which shows the structure of the test | inspection apparatus of this invention. It is a flowchart which shows a 1st test process and a 2nd test process. It is a flowchart which shows a 3rd test | inspection process. It is explanatory drawing which shows the test | inspection of the solid-state image sensor by the test | inspection apparatus of this invention. It is a flowchart which shows the 3rd inspection process in the case of performing the quality determination based on the number of peculiar pixels, and the quality determination based on a defect level. It is a flowchart which shows the 3rd test | inspection process in the case of determining quality from the allowable upper limit number and the number of specific pixels. It is a flowchart which shows the 3rd test | inspection process in case the quality determination based on a defect level is performed before the quality determination based on the number of peculiar pixels.

Explanation of symbols

10 Inspection device (Solid-state image sensor inspection device)
15 CCD (solid-state image sensor)
41 Inspection Unit 42 Singular Pixel Detection Unit (Singular Pixel Detection Unit)
43 isolated singular pixel removing unit 44 singular pixel characteristic calculating unit (defect level calculating means, singular pixel counting means)
45 Unique pixel number evaluation section (singular pixel number evaluation means)
46 Defect level evaluation unit (defect level evaluation means)
51 Image data 52 Inspection image data 53 Predetermined range 54, 56, 57, 58, 59 Dense singular pixels

Claims (16)

A first inspection step of acquiring image data in a state where a solid-state imaging device to be inspected is shielded from light;
A second inspection step of comparing the data output from each pixel of the solid-state imaging device with a threshold value based on the image data and detecting a specific pixel that outputs data having a value larger than the threshold value;
Inspecting the image data in units of a predetermined range, calculating a defect level that is a total value of data output from the singular pixels within the predetermined range, and comparing the defect level with a predetermined upper limit value. A solid-state image sensor inspection method, comprising: a third inspection step for determining pass / fail of the solid-state image sensor. In the third inspection step, the defect level is calculated and the number of the specific pixels within the range to be inspected is counted.
The pass / fail determination performed by comparing the defect level with a predetermined upper limit value and the pass / fail determination performed according to the number of specific pixels are combined to perform pass / fail determination of the solid-state imaging device. The solid-state image sensor inspection method according to claim 1. According to the result of the quality determination of the solid-state image sensor performed according to the number of the singular pixels, the quality determination of the solid-state image sensor performed by comparing the defect level with a predetermined upper limit value is performed. The solid-state imaging device inspection method according to claim 2, wherein: The solid-state imaging device, which is performed by comparing the defect level with a predetermined upper limit value when the solid-state imaging device is determined to be defective in the quality determination of the solid-state imaging device performed according to the number of singular pixels. The solid-state imaging device inspection method according to claim 3, wherein the quality determination is performed. The solid-state image sensor that is performed by comparing the defect level with a predetermined upper limit value when the solid-state image sensor is determined to be good in the quality determination of the solid-state image sensor performed according to the number of the singular pixels. The solid-state imaging device inspection method according to claim 3, wherein the quality determination is performed. According to the result of the quality determination of the solid-state image sensor performed by comparing the defect level with a predetermined upper limit value, the quality determination of the solid-state image sensor performed according to the number of the specific pixels is performed. The solid-state imaging device inspection method according to claim 2, wherein: In the second inspection step, the singular pixel is detected, and the data of the isolated singular pixel, which is a pixel in which all adjacent pixels output data having a value equal to or less than the threshold value, is removed from the image data. Inspection image data in which only the data of the singular pixel is left is created,
The solid-state imaging device inspection method according to claim 1, wherein the third inspection step is performed based on the inspection image data. 8. The solid-state imaging device inspection method according to claim 1, wherein the number of pixels included in the predetermined range is 0.001% to 0.01% of all pixels of the solid-state imaging device. . Based on the image data obtained by imaging the solid-state imaging device to be inspected in a light-shielded state, the data output from each pixel of the solid-state imaging device is compared with a threshold value, and data that is larger than the threshold value is output Singular pixel detection means for detecting pixels;
Defect level calculation means for calculating a defect level that is a total value of data output by the specific pixels within the predetermined range, with a predetermined range as a unit of the image data;
A solid-state image sensor inspection apparatus comprising: defect level evaluation means for performing pass / fail determination by comparing the defect level with a predetermined upper limit value. Singular pixel counting means for counting the number of the singular pixels within the predetermined range;
Singular pixel number evaluation means for performing pass / fail determination according to the number of singular pixels,
The solid-state imaging device inspection apparatus according to claim 9, wherein the quality determination of the solid-state imaging device is performed by combining the quality determination by the defect level evaluation unit and the quality determination by the specific pixel number evaluation unit. The solid-state image sensor inspection apparatus according to claim 10, wherein the pass / fail determination by the defect level evaluation unit is performed according to a result of the pass / fail determination by the singular pixel number evaluation unit.   12. The solid-state image pickup device according to claim 11, wherein the pass / fail judgment by the defect level evaluation unit is performed when the solid-state image pickup unit is determined to be defective as a result of the pass / fail determination by the singular pixel number evaluation unit. Inspection device.   12. The solid-state image pickup device according to claim 11, wherein the pass / fail judgment by the defect level evaluation unit is performed when the solid-state image pickup unit is determined to be good as a pass / fail judgment defect by the singular pixel number evaluation unit. Inspection device.   The solid-state image sensor inspection device according to claim 10, wherein the quality determination by the singular pixel number evaluation unit is performed according to a result of the quality determination by the defect level evaluation unit.   A solid-state image sensor, which is inspected by the solid-state image sensor inspection method according to claim 1.   A solid-state image sensor, which is inspected by the solid-state image sensor inspection apparatus according to claim 9.

Images